Popular websites such as Yelp use recommender systems to match users’ preferences with relevant services. While these systems make recommendations based on basic questions, such as “What are you in the mood for?”, they are unable to consider deeper context that influences the purchasing decisions of consumers.

To overcome this limitation, Konstantin Bauman developed a systematic method of extracting meaningful contextual information from online reviews. His method identifies key phrases that capture pieces of context most closely related to the reviewer’s purchase and service ratings.

This method can be used to help existing recommender systems ask smarter questions, make better suggestions and improve consumer experiences with service providers.